An analysis of the Max-Min approach to feature selection and ordering

The causes of deficiency of the Max-Min approach to feature selection are investigated. It is shown that the theoretical premise providing the basis for the Max-Min algorithm is not necessarily valid. A condition under which the Max-Min algorithm is not justified is derived, and a counterexample ill...

Full description

Saved in:
Bibliographic Details
Published inPattern recognition letters Vol. 14; no. 11; pp. 841 - 847
Main Authors Pudil, P., Novovičová, J., Choakjarernwanit, N., Kittler, J.
Format Journal Article
LanguageEnglish
Published Amsterdam Elsevier B.V 01.11.1993
Elsevier
Subjects
Applied sciences
Artificial intelligence
Computer science; control theory; systems
Exact sciences and technology
feature ordering
Feature selection
Max-Min algorithm
Pattern recognition. Digital image processing. Computational geometry
Max-Min algorithm
Feature selection
feature ordering
Feature ordering
Pattern recognition
Online AccessGet full text

Cover

More Information
Summary:The causes of deficiency of the Max-Min approach to feature selection are investigated. It is shown that the theoretical premise providing the basis for the Max-Min algorithm is not necessarily valid. A condition under which the Max-Min algorithm is not justified is derived, and a counterexample illustrating it is presented.
ISSN:0167-8655
1872-7344
DOI:10.1016/0167-8655(93)90147-6